Face Generation¶

In this project, you'll use generative adversarial networks to generate new images of faces.

Get the Data¶

You'll be using two datasets in this project:

MNIST
CelebA

Since the celebA dataset is complex and you're doing GANs in a project for the first time, we want you to test your neural network on MNIST before CelebA. Running the GANs on MNIST will allow you to see how well your model trains sooner.

If you're using FloydHub, set data_dir to "/input" and use the FloydHub data ID "R5KrjnANiKVhLWAkpXhNBe".

data_dir = './data'

# FloydHub - Use with data ID "R5KrjnANiKVhLWAkpXhNBe"
#data_dir = '/input'


"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import helper

helper.download_extract('mnist', data_dir)
helper.download_extract('celeba', data_dir)

Found mnist Data
Found celeba Data

Explore the Data¶

MNIST¶

As you're aware, the MNIST dataset contains images of handwritten digits. You can view the first number of examples by changing show_n_images.

show_n_images = 25

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
%matplotlib inline
import os
from glob import glob
from matplotlib import pyplot

mnist_images = helper.get_batch(glob(os.path.join(data_dir, 'mnist/*.jpg'))[:show_n_images], 28, 28, 'L')
pyplot.imshow(helper.images_square_grid(mnist_images, 'L'), cmap='gray')

<matplotlib.image.AxesImage at 0x7fc550502198>

CelebA¶

The CelebFaces Attributes Dataset (CelebA) dataset contains over 200,000 celebrity images with annotations. Since you're going to be generating faces, you won't need the annotations. You can view the first number of examples by changing show_n_images.

show_n_images = 25

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
mnist_images = helper.get_batch(glob(os.path.join(data_dir, 'img_align_celeba/*.jpg'))[:show_n_images], 28, 28, 'RGB')
pyplot.imshow(helper.images_square_grid(mnist_images, 'RGB'))

<matplotlib.image.AxesImage at 0x7fc567e632b0>

Preprocess the Data¶

Since the project's main focus is on building the GANs, we'll preprocess the data for you. The values of the MNIST and CelebA dataset will be in the range of -0.5 to 0.5 of 28x28 dimensional images. The CelebA images will be cropped to remove parts of the image that don't include a face, then resized down to 28x28.

The MNIST images are black and white images with a single color channel while the CelebA images have 3 color channels (RGB color channel).

Build the Neural Network¶

You'll build the components necessary to build a GANs by implementing the following functions below:

model_inputs
discriminator
generator
model_loss
model_opt
train

Check the Version of TensorFlow and Access to GPU¶

This will check to make sure you have the correct version of TensorFlow and access to a GPU

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
from distutils.version import LooseVersion
import warnings
import tensorflow as tf

# Check TensorFlow Version
assert LooseVersion(tf.__version__) >= LooseVersion('1.0'), 'Please use TensorFlow version 1.0 or newer.  You are using {}'.format(tf.__version__)
print('TensorFlow Version: {}'.format(tf.__version__))

# Check for a GPU
if not tf.test.gpu_device_name():
    warnings.warn('No GPU found. Please use a GPU to train your neural network.')
else:
    print('Default GPU Device: {}'.format(tf.test.gpu_device_name()))

TensorFlow Version: 1.0.1
Default GPU Device: /gpu:0

Input¶

Implement the model_inputs function to create TF Placeholders for the Neural Network. It should create the following placeholders:

Real input images placeholder with rank 4 using image_width, image_height, and image_channels.
Z input placeholder with rank 2 using z_dim.
Learning rate placeholder with rank 0.

Return the placeholders in the following the tuple (tensor of real input images, tensor of z data)

import problem_unittests as tests

def model_inputs(image_width, image_height, image_channels, z_dim):
    """
    Create the model inputs
    :param image_width: The input image width
    :param image_height: The input image height
    :param image_channels: The number of image channels
    :param z_dim: The dimension of Z
    :return: Tuple of (tensor of real input images, tensor of z data, learning rate)
    """
    # TODO: Implement Function
    real_image = tf.placeholder(tf.float32, (None, image_width, image_height, image_channels), name = "real_image")
    z_input = tf.placeholder(tf.float32, (None, z_dim), name="z_input")
    learning_rate = tf.placeholder(tf.float32, name="learning_rate")
    return real_image, z_input, learning_rate


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_inputs(model_inputs)

Tests Passed

Discriminator¶

Implement discriminator to create a discriminator neural network that discriminates on images. This function should be able to reuse the variabes in the neural network. Use tf.variable_scope with a scope name of "discriminator" to allow the variables to be reused. The function should return a tuple of (tensor output of the discriminator, tensor logits of the discriminator).

def discriminator(images, reuse=False):
    """
    Create the discriminator network
    :param image: Tensor of input image(s)
    :param reuse: Boolean if the weights should be reused
    :return: Tuple of (tensor output of the discriminator, tensor logits of the discriminator)
    """
    # TODO: Implement Function
    with tf.variable_scope('discriminator', reuse=reuse):
        
        alpha = 0.2
        
        #shape of image 28X28X1 in mnsit and 28X28X3 in case of celeb images
        x = tf.layers.conv2d(images, 64, 5, strides=2, padding="same")
        
        x1 = tf.layers.conv2d(x, 128, 5, strides=2, padding="same")
        x1 = tf.layers.batch_normalization(x1, training=True)
        x1 = tf.maximum(x1 * alpha, x1)
        
        x2 = tf.layers.conv2d(x1, 256, 5, strides=2, padding="same")
        x2 = tf.layers.batch_normalization(x2, training=True)
        x2 = tf.maximum(x2 * alpha, x2)
        
        shape = x2.get_shape().as_list()
        resshaped = tf.reshape(x2, (-1, shape[1] * shape[2] * shape[3]))
        logits = tf.layers.dense(resshaped, 1)
        output = tf.sigmoid(logits)
        return output, logits


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_discriminator(discriminator, tf)

Tests Passed

Generator¶

Implement generator to generate an image using z. This function should be able to reuse the variabes in the neural network. Use tf.variable_scope with a scope name of "generator" to allow the variables to be reused. The function should return the generated 28 x 28 x out_channel_dim images.

def generator(z, out_channel_dim, is_train=True):
    """
    Create the generator network
    :param z: Input z
    :param out_channel_dim: The number of channels in the output image
    :param is_train: Boolean if generator is being used for training
    :return: The tensor output of the generator
    """
    # TODO: Implement Function
    with tf.variable_scope('generator', reuse=(not is_train)):
        alpha = 0.2
        x = tf.layers.dense(z, 7*7*256)
        reshaped = tf.reshape(x, (-1, 7, 7, 256))
        x1 = tf.layers.batch_normalization(reshaped, training=is_train)
        x1 = tf.maximum(x1 * alpha, x1)
        
        x2 = tf.layers.conv2d_transpose(x1, 128, 5, strides=2, padding="same")
        x2 = tf.layers.batch_normalization(x2, training=is_train)
        x2 = tf.maximum(x2 * alpha , x2)
        
#         x3 = tf.layers.conv2d_transpose(x2, 128, 5, strides=2, padding="same")
#         x3 = tf.layers.batch_normalization(x3, training=is_train)
#         x3 = tf.maximum(x3 * alpha, x3)
        
        logits = tf.layers.conv2d_transpose(x2, out_channel_dim, 5, strides=2, padding="same")
        output = tf.tanh(logits)
        return output


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_generator(generator, tf)

Tests Passed

Loss¶

Implement model_loss to build the GANs for training and calculate the loss. The function should return a tuple of (discriminator loss, generator loss). Use the following functions you implemented:

discriminator(images, reuse=False)
generator(z, out_channel_dim, is_train=True)

def model_loss(input_real, input_z, out_channel_dim):
    """
    Get the loss for the discriminator and generator
    :param input_real: Images from the real dataset
    :param input_z: Z input
    :param out_channel_dim: The number of channels in the output image
    :return: A tuple of (discriminator loss, generator loss)
    """
    # TODO: Implement Function
    g_model = generator(input_z, out_channel_dim, is_train=True)
    d_model_real, d_logits_real = discriminator(input_real)
    d_model_fake, d_logits_fake = discriminator(g_model, reuse=True)
    
    d_loss_real = tf.reduce_mean(
        tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_real, labels=tf.ones_like(d_model_real)))
    d_loss_fake = tf.reduce_mean(
        tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_fake, labels=tf.zeros_like(d_model_fake)))
    g_loss = tf.reduce_mean(
        tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_fake, labels=tf.ones_like(d_model_fake)))

    d_loss = d_loss_real + d_loss_fake
    return d_loss, g_loss


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_loss(model_loss)

Tests Passed

Optimization¶

Implement model_opt to create the optimization operations for the GANs. Use tf.trainable_variables to get all the trainable variables. Filter the variables with names that are in the discriminator and generator scope names. The function should return a tuple of (discriminator training operation, generator training operation).

def model_opt(d_loss, g_loss, learning_rate, beta1):
    """
    Get optimization operations
    :param d_loss: Discriminator loss Tensor
    :param g_loss: Generator loss Tensor
    :param learning_rate: Learning Rate Placeholder
    :param beta1: The exponential decay rate for the 1st moment in the optimizer
    :return: A tuple of (discriminator training operation, generator training operation)
    """
    # TODO: Implement Function
    t_vars = tf.trainable_variables()
    d_vars = [var for var in t_vars if var.name.startswith('discriminator')]
    g_vars = [var for var in t_vars if var.name.startswith('generator')]

    # Optimize
    with tf.control_dependencies(tf.get_collection(tf.GraphKeys.UPDATE_OPS)):
        d_train_opt = tf.train.AdamOptimizer(learning_rate, beta1=beta1).minimize(d_loss, var_list=d_vars)
        g_train_opt = tf.train.AdamOptimizer(learning_rate, beta1=beta1).minimize(g_loss, var_list=g_vars)
        return d_train_opt, g_train_opt


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_opt(model_opt, tf)

Tests Passed

Neural Network Training¶

Show Output¶

Use this function to show the current output of the generator during training. It will help you determine how well the GANs is training.

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import numpy as np

def show_generator_output(sess, n_images, input_z, out_channel_dim, image_mode):
    """
    Show example output for the generator
    :param sess: TensorFlow session
    :param n_images: Number of Images to display
    :param input_z: Input Z Tensor
    :param out_channel_dim: The number of channels in the output image
    :param image_mode: The mode to use for images ("RGB" or "L")
    """
    cmap = None if image_mode == 'RGB' else 'gray'
    z_dim = input_z.get_shape().as_list()[-1]
    example_z = np.random.uniform(-1, 1, size=[n_images, z_dim])

    samples = sess.run(
        generator(input_z, out_channel_dim, False),
        feed_dict={input_z: example_z})

    images_grid = helper.images_square_grid(samples, image_mode)
    pyplot.imshow(images_grid, cmap=cmap)
    pyplot.show()

Train¶

Implement train to build and train the GANs. Use the following functions you implemented:

model_inputs(image_width, image_height, image_channels, z_dim)
model_loss(input_real, input_z, out_channel_dim)
model_opt(d_loss, g_loss, learning_rate, beta1)

Use the show_generator_output to show generator output while you train. Running show_generator_output for every batch will drastically increase training time and increase the size of the notebook. It's recommended to print the generator output every 100 batches.

def train(epoch_count, batch_size, z_dim, learning_rate, beta1, get_batches, data_shape, data_image_mode):
    """
    Train the GAN
    :param epoch_count: Number of epochs
    :param batch_size: Batch Size
    :param z_dim: Z dimension
    :param learning_rate: Learning Rate
    :param beta1: The exponential decay rate for the 1st moment in the optimizer
    :param get_batches: Function to get batches
    :param data_shape: Shape of the data
    :param data_image_mode: The image mode to use for images ("RGB" or "L")
    """
    # TODO: Build Model
    input_real, input_z, lr = model_inputs( data_shape[1], data_shape[2], data_shape[3], z_dim)
    d_loss, g_loss = model_loss(input_real, input_z, data_shape[3])
    d_opt, g_opt = model_opt(d_loss, g_loss, lr, beta1)
    
    samples, losses = [], []
    steps = 0
    
    with tf.Session() as sess:
        sess.run(tf.global_variables_initializer())
        for epoch_i in range(epoch_count):
            for batch_images in get_batches(batch_size):
                # TODO: Train Model
                batch_images = batch_images*2.0
                steps += 1
                batch_z = np.random.uniform(-1, 1, size=(batch_size, z_dim))
                
                #optimize d_opt
                _ = sess.run(d_opt, feed_dict={input_real: batch_images, input_z: batch_z, lr: learning_rate})
                #optimize g_opt
                _ = sess.run(g_opt, feed_dict={input_z: batch_z, input_real: batch_images, lr: learning_rate})
                _ = sess.run(g_opt, feed_dict={input_z: batch_z, input_real: batch_images, lr: learning_rate})
                
                if steps % 10 == 0:
                    # At the end of each epoch, get the losses and print them out
                    train_loss_d = d_loss.eval({input_z: batch_z, input_real: batch_images})
                    train_loss_g = g_loss.eval({input_z: batch_z})
                    
                    print("Epoch {}/{}...".format(epoch_i, epochs),
                          "Discriminator Loss: {:.4f}...".format(train_loss_d),
                          "Generator Loss: {:.4f}".format(train_loss_g))
                    # Save losses to view after training
                    losses.append((train_loss_d, train_loss_g))
                    
                if steps % 100 == 0:
                    show_generator_output(sess,25, input_z, data_shape[3], data_image_mode)

        return losses, samples

MNIST¶

Test your GANs architecture on MNIST. After 2 epochs, the GANs should be able to generate images that look like handwritten digits. Make sure the loss of the generator is lower than the loss of the discriminator or close to 0.

batch_size = 64
z_dim = 100
learning_rate = 0.0002
beta1 = 0.4


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
epochs = 5

mnist_dataset = helper.Dataset('mnist', glob(os.path.join(data_dir, 'mnist/*.jpg')))
with tf.Graph().as_default():
    train(epochs, batch_size, z_dim, learning_rate, beta1, mnist_dataset.get_batches,
          mnist_dataset.shape, mnist_dataset.image_mode)

Epoch 0/5... Discriminator Loss: 2.8921... Generator Loss: 0.1726
Epoch 0/5... Discriminator Loss: 2.7275... Generator Loss: 0.1468
Epoch 0/5... Discriminator Loss: 2.0156... Generator Loss: 0.2159
Epoch 0/5... Discriminator Loss: 1.8430... Generator Loss: 0.3370
Epoch 0/5... Discriminator Loss: 1.6812... Generator Loss: 0.3140
Epoch 0/5... Discriminator Loss: 1.8759... Generator Loss: 0.2272
Epoch 0/5... Discriminator Loss: 1.5370... Generator Loss: 0.4659
Epoch 0/5... Discriminator Loss: 1.6254... Generator Loss: 1.0007
Epoch 0/5... Discriminator Loss: 1.5441... Generator Loss: 0.3451
Epoch 0/5... Discriminator Loss: 1.5144... Generator Loss: 0.4055

Epoch 0/5... Discriminator Loss: 1.3805... Generator Loss: 0.6088
Epoch 0/5... Discriminator Loss: 1.5011... Generator Loss: 1.0787
Epoch 0/5... Discriminator Loss: 1.3036... Generator Loss: 0.6984
Epoch 0/5... Discriminator Loss: 1.4278... Generator Loss: 0.3808
Epoch 0/5... Discriminator Loss: 1.2737... Generator Loss: 0.4931
Epoch 0/5... Discriminator Loss: 1.3180... Generator Loss: 0.7100
Epoch 0/5... Discriminator Loss: 1.2958... Generator Loss: 1.1347
Epoch 0/5... Discriminator Loss: 1.1967... Generator Loss: 1.0206
Epoch 0/5... Discriminator Loss: 1.2025... Generator Loss: 0.6659
Epoch 0/5... Discriminator Loss: 1.1857... Generator Loss: 1.2116

Epoch 0/5... Discriminator Loss: 1.3862... Generator Loss: 0.4172
Epoch 0/5... Discriminator Loss: 1.1090... Generator Loss: 0.6967
Epoch 0/5... Discriminator Loss: 1.3661... Generator Loss: 0.4000
Epoch 0/5... Discriminator Loss: 1.1844... Generator Loss: 0.5744
Epoch 0/5... Discriminator Loss: 1.0457... Generator Loss: 0.7745
Epoch 0/5... Discriminator Loss: 1.3399... Generator Loss: 0.4263
Epoch 0/5... Discriminator Loss: 1.2543... Generator Loss: 0.5012
Epoch 0/5... Discriminator Loss: 1.0209... Generator Loss: 0.7702
Epoch 0/5... Discriminator Loss: 1.0732... Generator Loss: 1.1665
Epoch 0/5... Discriminator Loss: 0.9991... Generator Loss: 0.8616

Epoch 0/5... Discriminator Loss: 1.4130... Generator Loss: 0.3624
Epoch 0/5... Discriminator Loss: 1.0845... Generator Loss: 0.6053
Epoch 0/5... Discriminator Loss: 1.3848... Generator Loss: 0.3822
Epoch 0/5... Discriminator Loss: 1.3526... Generator Loss: 0.4335
Epoch 0/5... Discriminator Loss: 1.4660... Generator Loss: 0.3735
Epoch 0/5... Discriminator Loss: 1.3093... Generator Loss: 0.4711
Epoch 0/5... Discriminator Loss: 1.2350... Generator Loss: 0.5385
Epoch 0/5... Discriminator Loss: 1.2443... Generator Loss: 0.5492
Epoch 0/5... Discriminator Loss: 1.1887... Generator Loss: 0.8724
Epoch 0/5... Discriminator Loss: 1.3300... Generator Loss: 0.4912

Epoch 0/5... Discriminator Loss: 1.2683... Generator Loss: 0.5382
Epoch 0/5... Discriminator Loss: 1.2955... Generator Loss: 0.6178
Epoch 0/5... Discriminator Loss: 1.3420... Generator Loss: 1.0332
Epoch 0/5... Discriminator Loss: 1.3028... Generator Loss: 0.5149
Epoch 0/5... Discriminator Loss: 1.4211... Generator Loss: 0.4369
Epoch 0/5... Discriminator Loss: 1.3041... Generator Loss: 0.9102
Epoch 0/5... Discriminator Loss: 1.4980... Generator Loss: 0.9957
Epoch 0/5... Discriminator Loss: 1.3353... Generator Loss: 0.5296
Epoch 0/5... Discriminator Loss: 1.1933... Generator Loss: 0.7697
Epoch 0/5... Discriminator Loss: 1.2931... Generator Loss: 0.5585

Epoch 0/5... Discriminator Loss: 1.1612... Generator Loss: 0.7534
Epoch 0/5... Discriminator Loss: 1.1851... Generator Loss: 0.7114
Epoch 0/5... Discriminator Loss: 1.3280... Generator Loss: 0.5405
Epoch 0/5... Discriminator Loss: 1.3932... Generator Loss: 1.1867
Epoch 0/5... Discriminator Loss: 1.3352... Generator Loss: 0.7294
Epoch 0/5... Discriminator Loss: 1.3012... Generator Loss: 0.9303
Epoch 0/5... Discriminator Loss: 1.3966... Generator Loss: 1.3318
Epoch 0/5... Discriminator Loss: 1.4019... Generator Loss: 0.4641
Epoch 0/5... Discriminator Loss: 1.1640... Generator Loss: 0.6907
Epoch 0/5... Discriminator Loss: 1.2713... Generator Loss: 0.9456

Epoch 0/5... Discriminator Loss: 1.5182... Generator Loss: 0.3599
Epoch 0/5... Discriminator Loss: 1.1682... Generator Loss: 0.8202
Epoch 0/5... Discriminator Loss: 1.2584... Generator Loss: 0.5966
Epoch 0/5... Discriminator Loss: 1.3637... Generator Loss: 0.4709
Epoch 0/5... Discriminator Loss: 1.3898... Generator Loss: 0.9694
Epoch 0/5... Discriminator Loss: 1.2095... Generator Loss: 0.7994
Epoch 0/5... Discriminator Loss: 1.2469... Generator Loss: 0.8271
Epoch 0/5... Discriminator Loss: 1.2751... Generator Loss: 0.5243
Epoch 0/5... Discriminator Loss: 1.2832... Generator Loss: 0.6743
Epoch 0/5... Discriminator Loss: 1.2598... Generator Loss: 0.5326

Epoch 0/5... Discriminator Loss: 1.2881... Generator Loss: 0.7086
Epoch 0/5... Discriminator Loss: 1.2408... Generator Loss: 0.7090
Epoch 0/5... Discriminator Loss: 1.4308... Generator Loss: 0.7510
Epoch 0/5... Discriminator Loss: 1.2458... Generator Loss: 0.6443
Epoch 0/5... Discriminator Loss: 1.5091... Generator Loss: 0.3434
Epoch 0/5... Discriminator Loss: 1.3063... Generator Loss: 0.7875
Epoch 0/5... Discriminator Loss: 1.2783... Generator Loss: 0.8132
Epoch 0/5... Discriminator Loss: 1.3054... Generator Loss: 0.7803
Epoch 0/5... Discriminator Loss: 1.2101... Generator Loss: 0.8793
Epoch 0/5... Discriminator Loss: 1.2945... Generator Loss: 0.7100

Epoch 0/5... Discriminator Loss: 1.2917... Generator Loss: 0.8261
Epoch 0/5... Discriminator Loss: 1.3459... Generator Loss: 0.6851
Epoch 0/5... Discriminator Loss: 1.3630... Generator Loss: 0.5686
Epoch 0/5... Discriminator Loss: 1.3803... Generator Loss: 0.4895
Epoch 0/5... Discriminator Loss: 1.2296... Generator Loss: 0.7821
Epoch 0/5... Discriminator Loss: 1.2015... Generator Loss: 0.7642
Epoch 0/5... Discriminator Loss: 1.3906... Generator Loss: 1.1224
Epoch 0/5... Discriminator Loss: 1.1614... Generator Loss: 0.6689
Epoch 0/5... Discriminator Loss: 1.3582... Generator Loss: 0.6075
Epoch 0/5... Discriminator Loss: 1.3340... Generator Loss: 0.5498

Epoch 0/5... Discriminator Loss: 1.4100... Generator Loss: 0.4176
Epoch 0/5... Discriminator Loss: 1.4764... Generator Loss: 0.3566
Epoch 0/5... Discriminator Loss: 1.3289... Generator Loss: 0.4525
Epoch 1/5... Discriminator Loss: 1.3157... Generator Loss: 0.6476
Epoch 1/5... Discriminator Loss: 1.2387... Generator Loss: 0.9488
Epoch 1/5... Discriminator Loss: 1.3467... Generator Loss: 0.5737
Epoch 1/5... Discriminator Loss: 1.1921... Generator Loss: 0.7834
Epoch 1/5... Discriminator Loss: 1.2294... Generator Loss: 0.6575
Epoch 1/5... Discriminator Loss: 1.3561... Generator Loss: 0.7652
Epoch 1/5... Discriminator Loss: 1.3073... Generator Loss: 0.4946

CelebA¶

Run your GANs on CelebA. It will take around 20 minutes on the average GPU to run one epoch. You can run the whole epoch or stop when it starts to generate realistic faces.

batch_size = 64
z_dim = 100
learning_rate = 0.0002
beta1 = 0.5


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
epochs = 5

celeba_dataset = helper.Dataset('celeba', glob(os.path.join(data_dir, 'img_align_celeba/*.jpg')))
with tf.Graph().as_default():
    train(epochs, batch_size, z_dim, learning_rate, beta1, celeba_dataset.get_batches,
          celeba_dataset.shape, celeba_dataset.image_mode)

Epoch 0/5... Discriminator Loss: 2.1545... Generator Loss: 0.2936
Epoch 0/5... Discriminator Loss: 1.8445... Generator Loss: 0.5786
Epoch 0/5... Discriminator Loss: 2.1359... Generator Loss: 1.8131
Epoch 0/5... Discriminator Loss: 1.7407... Generator Loss: 0.4551
Epoch 0/5... Discriminator Loss: 1.8327... Generator Loss: 0.3552
Epoch 0/5... Discriminator Loss: 1.8071... Generator Loss: 0.3971
Epoch 0/5... Discriminator Loss: 1.9325... Generator Loss: 0.6375
Epoch 0/5... Discriminator Loss: 1.7798... Generator Loss: 0.3760
Epoch 0/5... Discriminator Loss: 1.8227... Generator Loss: 0.7007
Epoch 0/5... Discriminator Loss: 1.7205... Generator Loss: 0.4995

Epoch 0/5... Discriminator Loss: 1.4664... Generator Loss: 0.6297
Epoch 0/5... Discriminator Loss: 1.5445... Generator Loss: 0.4283
Epoch 0/5... Discriminator Loss: 1.6729... Generator Loss: 0.7443
Epoch 0/5... Discriminator Loss: 1.8423... Generator Loss: 0.5602
Epoch 0/5... Discriminator Loss: 1.6045... Generator Loss: 0.6764
Epoch 0/5... Discriminator Loss: 1.7521... Generator Loss: 0.4766
Epoch 0/5... Discriminator Loss: 1.6584... Generator Loss: 0.6871
Epoch 0/5... Discriminator Loss: 1.6098... Generator Loss: 0.9138
Epoch 0/5... Discriminator Loss: 1.5806... Generator Loss: 0.5819
Epoch 0/5... Discriminator Loss: 1.4714... Generator Loss: 0.6535

Epoch 0/5... Discriminator Loss: 1.5872... Generator Loss: 0.5586
Epoch 0/5... Discriminator Loss: 1.4920... Generator Loss: 0.6251
Epoch 0/5... Discriminator Loss: 1.5713... Generator Loss: 0.5870
Epoch 0/5... Discriminator Loss: 1.5896... Generator Loss: 0.4945
Epoch 0/5... Discriminator Loss: 1.4261... Generator Loss: 0.6547
Epoch 0/5... Discriminator Loss: 1.3829... Generator Loss: 0.6237
Epoch 0/5... Discriminator Loss: 1.5321... Generator Loss: 0.5686
Epoch 0/5... Discriminator Loss: 1.4667... Generator Loss: 0.6011
Epoch 0/5... Discriminator Loss: 1.5794... Generator Loss: 0.5691
Epoch 0/5... Discriminator Loss: 1.4016... Generator Loss: 0.6378

Epoch 0/5... Discriminator Loss: 1.3787... Generator Loss: 0.6536
Epoch 0/5... Discriminator Loss: 1.5339... Generator Loss: 0.6609
Epoch 0/5... Discriminator Loss: 1.4166... Generator Loss: 0.5386
Epoch 0/5... Discriminator Loss: 1.4935... Generator Loss: 0.6282
Epoch 0/5... Discriminator Loss: 1.5695... Generator Loss: 0.4758
Epoch 0/5... Discriminator Loss: 1.5896... Generator Loss: 0.5559
Epoch 0/5... Discriminator Loss: 1.5244... Generator Loss: 0.5186
Epoch 0/5... Discriminator Loss: 1.5516... Generator Loss: 0.5948
Epoch 0/5... Discriminator Loss: 1.4168... Generator Loss: 0.6404
Epoch 0/5... Discriminator Loss: 1.5287... Generator Loss: 0.5390

Epoch 0/5... Discriminator Loss: 1.5384... Generator Loss: 0.6342
Epoch 0/5... Discriminator Loss: 1.6456... Generator Loss: 0.5461
Epoch 0/5... Discriminator Loss: 1.6076... Generator Loss: 0.4994
Epoch 0/5... Discriminator Loss: 1.6863... Generator Loss: 0.4639
Epoch 0/5... Discriminator Loss: 1.5281... Generator Loss: 0.5381
Epoch 0/5... Discriminator Loss: 1.6706... Generator Loss: 0.5492
Epoch 0/5... Discriminator Loss: 1.5182... Generator Loss: 0.6083
Epoch 0/5... Discriminator Loss: 1.6584... Generator Loss: 0.4451
Epoch 0/5... Discriminator Loss: 1.5080... Generator Loss: 0.6581
Epoch 0/5... Discriminator Loss: 1.6109... Generator Loss: 0.6876

Epoch 0/5... Discriminator Loss: 1.4209... Generator Loss: 0.6706
Epoch 0/5... Discriminator Loss: 1.6033... Generator Loss: 0.5416
Epoch 0/5... Discriminator Loss: 1.5826... Generator Loss: 0.5180
Epoch 0/5... Discriminator Loss: 1.5533... Generator Loss: 0.5885
Epoch 0/5... Discriminator Loss: 1.4508... Generator Loss: 0.6424
Epoch 0/5... Discriminator Loss: 1.5154... Generator Loss: 0.5834
Epoch 0/5... Discriminator Loss: 1.5792... Generator Loss: 0.6138
Epoch 0/5... Discriminator Loss: 1.4754... Generator Loss: 0.7061
Epoch 0/5... Discriminator Loss: 1.5227... Generator Loss: 0.6112
Epoch 0/5... Discriminator Loss: 1.6034... Generator Loss: 0.5765

Epoch 0/5... Discriminator Loss: 1.5249... Generator Loss: 0.6577
Epoch 0/5... Discriminator Loss: 1.4984... Generator Loss: 0.6489
Epoch 0/5... Discriminator Loss: 1.5286... Generator Loss: 0.5717
Epoch 0/5... Discriminator Loss: 1.4576... Generator Loss: 0.7176
Epoch 0/5... Discriminator Loss: 1.4744... Generator Loss: 0.6103
Epoch 0/5... Discriminator Loss: 1.5015... Generator Loss: 0.6559
Epoch 0/5... Discriminator Loss: 1.4314... Generator Loss: 0.6483
Epoch 0/5... Discriminator Loss: 1.4883... Generator Loss: 0.6625
Epoch 0/5... Discriminator Loss: 1.4614... Generator Loss: 0.7011
Epoch 0/5... Discriminator Loss: 1.5683... Generator Loss: 0.6006

Epoch 0/5... Discriminator Loss: 1.4904... Generator Loss: 0.6262
Epoch 0/5... Discriminator Loss: 1.4935... Generator Loss: 0.6262
Epoch 0/5... Discriminator Loss: 1.4174... Generator Loss: 0.6185
Epoch 0/5... Discriminator Loss: 1.5601... Generator Loss: 0.6169
Epoch 0/5... Discriminator Loss: 1.4090... Generator Loss: 0.6271
Epoch 0/5... Discriminator Loss: 1.4855... Generator Loss: 0.6584
Epoch 0/5... Discriminator Loss: 1.5533... Generator Loss: 0.5798
Epoch 0/5... Discriminator Loss: 1.4310... Generator Loss: 0.7023
Epoch 0/5... Discriminator Loss: 1.5444... Generator Loss: 0.5848
Epoch 0/5... Discriminator Loss: 1.4513... Generator Loss: 0.6360

Epoch 0/5... Discriminator Loss: 1.4304... Generator Loss: 0.6717
Epoch 0/5... Discriminator Loss: 1.4832... Generator Loss: 0.6288
Epoch 0/5... Discriminator Loss: 1.5294... Generator Loss: 0.5780
Epoch 0/5... Discriminator Loss: 1.3916... Generator Loss: 0.6646
Epoch 0/5... Discriminator Loss: 1.4375... Generator Loss: 0.6862
Epoch 0/5... Discriminator Loss: 1.4784... Generator Loss: 0.6893
Epoch 0/5... Discriminator Loss: 1.4623... Generator Loss: 0.5971
Epoch 0/5... Discriminator Loss: 1.4604... Generator Loss: 0.6583
Epoch 0/5... Discriminator Loss: 1.4859... Generator Loss: 0.6101
Epoch 0/5... Discriminator Loss: 1.3521... Generator Loss: 0.6986

Epoch 0/5... Discriminator Loss: 1.5771... Generator Loss: 0.5780
Epoch 0/5... Discriminator Loss: 1.4855... Generator Loss: 0.6364
Epoch 0/5... Discriminator Loss: 1.4637... Generator Loss: 0.6260
Epoch 0/5... Discriminator Loss: 1.4736... Generator Loss: 0.6601
Epoch 0/5... Discriminator Loss: 1.5384... Generator Loss: 0.6372
Epoch 0/5... Discriminator Loss: 1.4046... Generator Loss: 0.6692
Epoch 0/5... Discriminator Loss: 1.5058... Generator Loss: 0.5991
Epoch 0/5... Discriminator Loss: 1.3988... Generator Loss: 0.6896
Epoch 0/5... Discriminator Loss: 1.4743... Generator Loss: 0.6471
Epoch 0/5... Discriminator Loss: 1.5062... Generator Loss: 0.6034

Submitting This Project¶

When submitting this project, make sure to run all the cells before saving the notebook. Save the notebook file as "dlnd_face_generation.ipynb" and save it as a HTML file under "File" -> "Download as". Include the "helper.py" and "problem_unittests.py" files in your submission.